Bloomberg Law
March 8, 2018, 11:45 AM

Humans Wanted: Big Law Needs People to Make AI Work

Jasmine Ye Han
Jasmine Ye Han
Data Reporter

Law firms looking at artificial intelligence and other technology solutions can expect to make a big investment in the humans to implement it.

Driven by “NextGen” legal technologies--which include artificial Intelligence, machine learning, and cognitive computing--new roles are emerging within law firms, said David Cowen, president of the Cowen Group, which specializes in filling legal technology and executive positions in law firms. The new roles include chief innovation officer, legal solutions architect, and chief data scientist, Cowen told Bloomberg Law.

“The technology looks incredible but it’s not just a push button thing to implement and deploy. It requires some expertise to get the tools to do what you want to achieve,” said Ruth Hauswirth, special counsel and director of litigation e-discovery services for Cooley. Hauswirth is leading a dedicated department of specialists focused on issues that data present in litigation.

These NextGen technologies are used to build tools to aid lawyers in e-discovery, document review, legal research, and beyond. For example, Ross Intelligence is a legal research tool that can answer questions in natural language entered by lawyers. LawGeex, a contract review platform, says it can spot risks more accurately and efficiently than human lawyers.

Human Expertise Necessary in Process Design and Use

Some see the emerging technology as a threat and are concerned that robots will replace lawyers. But Stephen Poor, chair emeritus of Seyfarth Shaw LLP, dismissed that notion.

“It’s important to think about automation not as replacing jobs, but replacing tasks; not replacing humans, but augmenting humans,” he told Bloomberg Law. Poor serves as an executive sponsor of strategic initiatives focused on innovation at his firm.

It takes deep and clear understanding of the process, and the role of technology and humans, to make the most out of the technology, Poor said. It also takes a willingness to redesign the process, because “if you have a bad process and use an A.I. that does the same thing, you’re just making it faster, not more efficient,” he said.

After the process is designed, human input is also necessary to use the tools properly, although how much depends on the system, Hauswirth said. In e-discovery technologies, which help identify documents relevant to a case from a huge amount of documents, lawyers are needed to provide information about what they’re looking for. Some systems can be “trained” by human lawyers. For example, they learn what’s relevant from a lawyer’s review of some documents and predict how the lawyers may decide for the rest of the body of documents.

Training of the systems is an important step to showcase the best legal language each law firm has developed over the years around certain provisions, “because ultimately, that’s the experience and expertise that cannot be duplicated by machine,” said Ali Shahidi, chief innovation officer at Sheppard Mullin Richter & Hampton LLP.

Testing, Tailoring Crucial to Adoption

Shahidi said he’s seeing an increasing appetite in law firms for experimentation, testing, evaluating the technologies, a “culture of innovation,” with firms ultimately rejecting some tools and adopting others.

BakerHostetler, for example, came up with “Project Black Box,” a sophisticated framework for testing and studying attorney use and adoption of technology tools, Katherine Lowry, leader of the firm’s practice services group, told Bloomberg Law. Her team is responsible for monitoring technology advancements with an eye toward leveraging those that increase efficiency, save cost, and reduce risks.

The idea of the project is to measure the trust gap between lawyers and emerging technology, which can keep them from capturing the true value of the tools. To provide a more complete picture of attorneys’ reaction to new technology, the approach uses both qualitative and quantitative data, Lowry said.

Lowry’s team uses the results generated by the testing to come up with recommendations, such as how vendors can improve certain features to increase usability. “We think the time invested in doing this saves us cost, because it ultimately allows us to choose the tools that fit our needs,” Lowry said.

Worth the Fuss?

The tools, if you include the attorney time and client engagement involved in rolling them out, aren’t inexpensive. “It takes a big investment up front,” Shahidi said.

Poor agrees but said his firm has seen a positive return on all these investments. “If you do it right, you wind up getting two things: lower cost, and happier clients. How much is that worth”? he asked.

Hauswirth said she sees the tools helping lawyers do reviews much more efficiently than they could in the past. And sometimes the tools open up a window to insights that attorneys otherwise couldn’t get.

“If you are talking about terabytes of data, there is no way to know what’s in there until you use these tools and have them help you learn what are the relationships between different data that you otherwise can’t see,” she said.

There’s certainly evidence that some of the tools can save time and cost, but it’s too early to tell about others, Lowry said. “But if you don’t invest time in them now, you won’t be ready for the changes coming two to five years from now,” she said.

To contact the reporter on this story: Jasmine Ye Han in Washington at

To contact the editor responsible for this story: Terence Hyland at; Jo-el J. Meyer at